Simulation of biomass gasification with a hybrid neural network model

Gasification experiments of several types of biomass were conducted in an atmospheric steam fluidized bed gasifier, which biomass samples were fed into continuously and without residue discharge. In order to obtain the gasification profiles for each type of biomass, an artificial neural network model has been developed to simulate the gasification process. Model-predicted gas production rates for the biomass gasification processes are in good agreement with the experimental data, thereby the gasification profiles given by the neural network model are considered to properly reflect the real gasification process of a biomass. Gasification profiles identified by neural network model suggest that gasification behavior of arboreal types of biomass is significantly different from that of herbaceous ones.